Daniel Kales
Daniel Kales
https://github.com/NilFoundation/zkllvm-blueprint/blob/93b6463d8d6378e08994c3208fbed34f90b30fa7/test/test_plonk_component.hpp#L286C38-L286C38 The component tester assumes that all components use at most 1 column for constants. We have components that in some configurations where the component fits into a single row,...
The CudaSlice abstraction does not have an API to split it into two mutable `CudaViewMut`s, which kind of forces a user to do so with `CudaView`s, and passing references to...
This is an attempt to address #233. Since this is code that touches some aspects of the lifetimes of `CudaView` and `CudaViewMut`, this could use some extra scrutiny to check...
During our integration of the lookup constraints in our components, we came across two questions. 1. Currently, the first row of all columns in the lookup tables is reserved for...
We are currently working to integrate lookups in our components We have several different lookup tables, where the first column is a simple 0..N index table and the second column...
In https://github.com/coreylowman/cudarc/blob/886d6d27cd68da4f81ce30a98bdf1940a895f813/src/nccl/safe.rs#L242-L245, the sendbuf is given as an `&Option`, which involves wrapping the `T`, which is something like a `CudaSlice` in an Option to pass it in there, making re-use...
We recently upgraded to `ark-* 0.5` and observed noticeable performance degradation on deserialization of large `Vec`. This is due to the compiler no longer being able to optimize out the...
First stab at adding intra-batch distances to anon stats. As a slight complication, like the matching itself, anon stats for intra-batch matches are only added if the matching element would...
This is in preperation for the fragile bits effort for iris codes. Up until now, mask code shares have had the same imaginary and real part, which meant that we...